Background To inform discussions on rates, burden and priority-setting in relation to police violence, we quantified the number and rate of years of life lost (YLLs) due to police violence by race/ethnicity and age in the USA, 2015–2016.
Methods We used data on the number of deaths due to police violence from ‘The Counted’, a media-based source compiled by The Guardian. YLLs are the difference between an individual’s age at death and their corresponding standard life expectancy at age of death.
Results There were 57 375 and 54 754 YLLs due to police violence in 2015 and 2016, respectively. People of colour comprised 38.5% of the population, but 51.5% of YLLs. YLLs were greatest among those aged 25–34 years, and the number of YLLs at younger ages was greater among people of colour than whites.
Conclusions The number of YLLs due to police violence is substantial. YLLs highlight that police violence disproportionately impacts young people, and the young people affected are disproportionately people of colour. Framing police violence as an important cause of deaths among young adults provides another valuable lens to motivate prevention efforts.
Statistics from Altmetric.com
Recent high-profile deaths and the Black Lives Matter movement have intensified focus on police violence as an important public health issue. Quantifying the magnitude of deaths and injuries from encounters with law enforcement (‘police violence’) and the degree of racial/ethnic disparities has been central to current discussions. While comprehensive national data on deaths from police violence are not currently collected,1 several studies have reported on the number or rate of deaths or injuries due to police violence using vital statistics, healthcare utilisation records or databases compiled from news reports.1 2 However, none has estimated years of life lost (YLLs).
YLLs are a useful metric for comparing the burden of diseases and injuries across different populations and health conditions, and for setting priorities and agendas in health policy. Mortality rates and counts naturally increase with age, so they inherently emphasise deaths at the oldest ages. In contrast, YLLs measure premature mortality by weighting each death by the number of years the person could have lived, according to an established ideal life expectancy standard, if he or she had not died prematurely. Some public health researchers and practitioners argue that public health efforts should focus on preventing premature death and therefore YLLs, not mortality counts or rates, should be used to make decisions about allocating scarce resources.3 4 To inform discussions on rates, burden and priority-setting in relation to police violence,5 we quantify the number and rate of YLLs due to police violence overall and by race/ethnicity and age in the USA in 2015 and 2016.
Deaths due to police violence are undercounted in official vital statistics and crime records.2 However, previous epidemiological studies have found that systematically compiled news accounts are a reliable source of injury data.2 6 In particular, ‘The Counted’, a publicly available dataset compiled by The Guardian,7 was designed to provide comprehensive fatality data by combining police reports, stories from news outlets and other independent reporting systems with crowdsourced information to serve as a surveillance system tracking all deaths in the USA related to police violence. Data from ‘The Counted’ have been validated as a source for the number of deaths due to police violence in the USA2 6 and found to be more complete than the National Vital Statistics System.8 We used death data from ‘The Counted’ to estimate the number of YLLs for each person in the dataset. We followed established methods, subtracting the age of death from the corresponding age-specific life expectancy from the standard life table used to calculate YLLs in the Global Burden of Disease Study 2015.9 We aggregated the YLLs for each individual by age and race/ethnicity. Eleven deaths (0.5%) without reported ages at death were excluded from the calculation of YLLs.
We estimated the rate of YLLs using Census-based denominators.10 Denominators were the most recent population projections for the years 2015 and 2016, based on the 2010 decennial Census and updated using the American Community Survey. We assumed that the Hispanic/Latino race/ethnicity from ‘The Counted’ corresponded with the Census Hispanic/Latino (any race) group, and that white, black, Asian/Pacific Islander and Native American groups in ‘The Counted’ corresponded with the Census single-race non-Hispanic white, black/African-American, Asian/Native Hawaiian/other Pacific Islander and American Indian/Alaskan Native groups, respectively. Rates for Arab-American, other and unknown groups in ‘The Counted’ were not estimated because population counts for these groups are not collected by the Census. In addition to showing results by race/ethnicity, we present results for people of colour, which includes all Hispanic and all non-white non-Hispanic racial/ethnic groups. To achieve stable estimates by race/ethnicity and for clarity, we averaged counts and rates over 2015 and 2016. Counts and rates by year, which show similar patterns, are presented in the online supplementary appendix.
Supplementary file 1
There were 1146 and 1092 deaths due to police violence in 2015 and 2016, respectively. Among these deaths, 51.7% were among whites, 25.6% were among blacks and 16.9% were among Hispanics (table 1). The average death rate due to police violence was 3.5 per million persons. American Indian/Alaskan Natives had the highest death rate at 7.8 per million, while blacks, Hispanics and whites had rates of 7.2, 3.3 and 2.9 persons per million, respectively.
An estimated 57 375 and 54 754 YLLs were due to police violence in 2015 and 2016, respectively. People of colour comprised 38.5% of the population, but 51.5% of all YLLs in 2015–2016. YLLs from police violence were greatest among younger age groups across racial and ethnic groups, but the distribution of YLLs was higher among even younger ages in people of colour compared with whites (figure 1). Compared with whites, the rate of YLLs was greater among American Indian/Alaskan Natives, blacks and Hispanics, and lower among Asian/Pacific Islanders (table 1). Across age groups, the number of YLLs was greatest among individuals aged 25–34 years. The median age at death for every racial/ethnic group, except for other/unknown, was notably lower than that for whites (table 1).
Discussion and public health implications
This study uses YLLs to quantify premature mortality due to police violence. YLLs combine the advantages of using death counts and age of death as metrics by including information both on the number of people who have died and the amount of their life lost prematurely. YLLs can also be aggregated to compare across time, demographic groups and causes of death.
The number of YLLs due to police violence is substantial. The burden is similar in magnitude to those due to meningitis (50 166 YLLs) and maternal deaths (56 490 YLLs), and greater than those due to cyclist road injuries (38 478) and unintentional firearm injuries (40 752).9 Yet, many of these conditions receive more attention than police violence, in terms of grant funding, for example.11
Quantifying the burden of police violence is critical to mobilising adequate public health, legal and policy responses. YLLs extend the framing of police violence as a population health issue, beyond what death counts and rates provide. Existing literature emphasises either that rates of police violence victimisation are much greater among people of colour compared with whites,1 or that despite this difference in rates, the number of whites killed by police is still greater.5 Similarly, we found that YLL counts were greatest among whites but YLL rates were greatest among people of colour. However, in addition to being an issue characterised by extreme racial/ethnic inequality, YLLs also suggest that police violence is an age issue. Police violence disproportionately impacts young people, and the young people affected are disproportionately people of colour. These findings are consistent with YLL age patterns for violence more broadly.9 Framing police violence as an important cause of deaths among young adults (which implicitly also means people of colour) provides another valuable lens to motivate prevention efforts. We hope that calculating YLLs for police violence can serve as a tool for advocates, including public health, legal and political leaders, as they continue to bring attention to how police violence impacts morbidity and mortality in this country, especially among youth, young adults and people of colour.
While measuring YLLs shows that police violence is an important contributor to premature mortality, it does not capture burden of non-fatal injuries, long-term disability and indirect effects of police violence including trauma to families and social networks, stress-related health outcomes and government instability.1 12 The burden of police violence on society measured in this study is therefore likely underestimated. Additionally, racial/ethnic groups in ‘The Counted’ align imperfectly with those in the Census. In particular, the exclusion of Arab Americans from the numerators of rates may underestimate race/ethnicity-specific YLL rates, and the assignment of multiracial individuals in ‘The Counted’ to a single racial/ethnic group likely resulted in some misclassification. However, YLL counts are unaffected by the former. Moreover, the use of ‘The Counted’ overcomes significant under-reporting in other sources, and other compiled news databases are less comprehensive (eg, The Washington Post only includes firearm deaths). Inadequacies in existing data remain a major barrier to appropriately assessing police violence. Producing better evidence, which is critical to relieving the burden that police violence has placed on youth, people of colour and society, will require greater public investment and improved data collection.
What is already known on this subject
Official national data on deaths from police violence are unavailable; however, several published analyses have used a combination of vital statistics, healthcare utilisation records or databases independently compiled from news reports to describe the numbers and rates of injuries and deaths from police violence.
Existing studies often point out the racial disparities in deaths from police violence, particularly the disproportionate rates of death among blacks.
Several studies have pointed out the need for deaths from encounters with law enforcement to be addressed as a public health issue and have suggested steps such as adding these deaths to the set of nationally notifiable conditions reported through local health departments.
What this study adds
To inform the public health response to police violence, this study characterises deaths due to police violence using a public health metric (years of life lost or YLLs) to describe the health burden in both rates and counts by race and age in 2015–2016. YLLs highlight the younger ages at which deaths from encounters with law enforcement occur in people of colour compared with whites.
Using YLLs shows deaths from police violence to rival significant causes of death in the USA, including meningitis and maternal deaths, and surpassing deaths from cyclist road injuries.
ALB, MMC and ECM contributed equally.
Contributors All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design; acquisition, analysis and interpretation of the data; drafting of the manuscript; critical revision of the manuscript for important intellectual content; statistical analysis; administrative, technical and material support; and study supervision: all authors.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent Not required.
Ethics approval This study was deemed not to constitute human subjects research by the University of California, Berkeley Committee for the Protection of Human Subjects.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Data used in this study were obtained from public sources including the U.S. Census and ’The Counted' dataset from the newspaper, The Guardian.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.